Abstract
Socio-Cognitive Mapping is an observational method to collect information about people’s social networks in settings in which participant observers know each other well, for example in school settings. Compared to traditional self-report data, observation reports make it possible to include (anonymized) network information about people who do not participate. In a series of papers, Neal and colleagues have criticized the methodology of Socio-Cognitive Mapping studies. However, the criticisms do not pertain to the data but only to a specific analysis program, SCM4, that was used in about 80% of the reviewed studies. To document their critiques, the authors introduce a new analysis strategy intended to correct some of the problems identified, and combine this with a promising new Community Detection method. They compare their results to SCM4 results and find in random simulations that, when using criteria that are more restrictive, fewer groups and fewer group members are identified. I highlight the extent to which the critique of the program is only justified under restrictive conditions, explain that the backbone of the proposed method has been used before, list problems of analyses that their method does not overcome, and outline avenues for their solution.